SGE: the definition, and why Google renamed it

News — The surface that launched as SGE keeps drifting from its experimental roots. At Google I/O in May 2026, Google reported that AI Mode had passed one billion monthly users and began folding AI Overviews and AI Mode into a single AI search experience running on its Gemini 3.5 Flash model (Google, I/O 2026, May 2026).

If you have run into the letters SGE and could not tell whether they name something that still exists, you are not confused, you are just early to a rename. This page defines the term, explains why Google buried it, and shows what it left behind. For the wider discipline that grew up around these AI answers, our GEO agency pillar covers the ground in depth.

SGE: the one-sentence definition

SGE stands for Search Generative Experience: the name Google used for its experimental AI-written answer at the top of the results page, launched in Search Labs in 2023, which summarised several web pages into a short response and showed a panel of the sources it drew from.

Two things in that sentence still matter, even though the label is gone. Experimental, because SGE was explicitly a test you opted into, not a default. And generative, because a model wrote the answer rather than lifting it from one page. The word people forget is Experience: Google named it after the whole way of searching, not a single box, which is precisely why the name became awkward once the feature stopped being an experiment and started showing up for everyone.

Where SGE came from, and what happened to the name

Google introduced SGE inside Search Labs as an opt-in experiment, then graduated the feature to general availability and renamed it AI Overviews. The technology carried over; the "experimental" framing, and the SGE label with it, was dropped.

Follow the naming and the strategy becomes legible. A Labs experiment can be called an "Experience", because you chose to try it and Google was still learning what it should be. A feature rolling out to a couple of billion people needs a plainer, product-level name, and "AI Overviews" is that name: it says what the thing is, an AI overview of the topic, without promising an experiment. So SGE did not fail and it did not get replaced. It grew up and changed its name, the way a codename gives way to a product name at launch.

That is the single most useful thing to hold onto. When you read an article from the SGE era, mentally translate SGE to AI Overviews, then check whether the specifics still hold, because the surface has kept moving since the rename. Google's own description of the machinery, that the answer is grounded in retrieved pages and the pages it used are surfaced as links, sits in its Search Central documentation (Google Search Central, 2025), and that part has stayed remarkably stable across the name change.

SGE vs AI Overviews vs AI Mode

SGE and AI Overviews are the same surface at two points in its life: the summary above the classic links. AI Mode is a different thing, a separate conversational tab where you ask follow-up questions, and Google is now merging the two experiences.

People conflate all three, and the confusion is expensive because the strategy is not identical. The table below is the version I keep coming back to when a client asks which of these they should be worrying about.

The convergence is the part to watch. At I/O in May 2026, Google described flowing straight from an AI Overview into a back-and-forth in AI Mode, and reported AI Mode passing one billion monthly users just a year after launch (Google, I/O 2026, 2026). The clean line between "a summary" and "a chatbot" is blurring, which is a good reason to stop over-indexing on any one label, including SGE.

What "SGE" means for your visibility today

Practically, SGE is a keyword you optimise for the same way you optimise for AI Overviews: by being retrievable, extractable and credible. The surface concentrates on longer, question-shaped searches, which is exactly where most of the demand actually sits.

Here is where our own numbers get uncomfortable. Of the 5092 French keywords Cicero Studio has analyzed, 35% draw fewer than 100 searches a month (Cicero Studio internal data). The long tail is not a rounding error, it is the bulk of the map, and it is made of precisely the spelled-out, question-shaped queries that summon a generative answer. Most teams pour their effort into a handful of fat-head terms and then get summarised on the long tail they ignored, which is the tail the SGE surface reads best.

Google is blunt about eligibility, and it is worth repeating because it kills a lot of magical thinking. To be shown as a supporting link, a page must be indexed and eligible to appear in Search with a snippet (Google Search Central, 2025). Block the snippet with nosnippet or an over-tight max-snippet and you have quietly excluded yourself from the surface, no matter how good the page is. There is no SGE-specific tag to add, and there never was.

What the surface changed for clicks

Clicks fall when a summary is present. Pew Research Center tracked 900 US adults and found users clicked a traditional result in 8% of visits with an AI summary, against 15% without, while clicks on the summary's own source links happened in just 1% of visits.

Nearly half the clicks, gone (Pew Research Center, 2025). Google counters that clicks arriving from pages shown with these AI answers tend to be higher quality, meaning the people who do click stay longer. Both statements can be true at once, and probably are: fewer visitors, better ones. But notice what neither side disputes. The visit is no longer the unit of visibility. Being named in the answer is.

That reframing is what our audit work keeps running into. Across the 1216 SEO and GEO audits Cicero Studio has run (Cicero Studio internal data), the recurring problem is almost never a shortage of content. It is content the machines cannot cleanly quote or confidently trust: a business publishes plenty, and the machine reader still cannot use what it publishes. That gap never shows up in a rankings report, which is exactly why it goes unnoticed for so long.

How to be present in the surface SGE became

Since Google offers no special optimisation, presence comes from three plain conditions: be retrievable, be extractable, and be credible. The research on generative engines found that cited statistics and named sources raise visibility inside AI answers, while keyword stuffing does nothing.

That finding, from the original academic work on generative engine optimisation (Aggarwal et al., arXiv, 2024), is the least glamorous and most useful thing in this field. The machine rewards what a good editor always rewarded: answer the question early, say something specific, name your source. If you want the cross-engine playbook rather than the principle, our practical method for appearing in ChatGPT and Google AI lays it out, and the current mechanics of the summary live on our AI Overviews definition. If the vocabulary is still fuzzy, start with the GEO definition or the AI Mode definition.

Alexis Dollé

CEO & Founder of Cicero Studio

I test AI visibility the slow way, by hand, one real business question at a time, across hundreds of sites. Names change, SGE to AI Overviews and on again, but the pages a Google AI answer cites are rarely the loudest ones. They are the clearest: a plain answer, stated early, backed by a source you can check. That is the standard we hold every page to at Cicero Studio.

LinkedIn →

What this page does not cover

A definition is only as useful as its edges, so here is what I left out, and where each thing lives instead.

Going further

We publish our method in the open, because that is the only proof worth much in this business.